Cameron Cooper

📍 New York, NY 📞 (917) 912-1264 ✉️ czc@umich.edu

Data Science and Computer Science student at the University of Michigan (B.S. Information, Data Science; Minor in Computer Science). I build AI systems that turn unstructured data into decisions. From deploying production NLP pipelines processing real estate portfolios at scale to creating full-stack quantitative research platforms, I focus on practical, high-impact tooling that bridges machine learning, software engineering, and business value.

Education

University of Michigan

Bachelor of Science in Data Science · Minor in Computer Science

GPA: 3.7

Relevant Coursework: Elementary Programming Concepts, Programming and Intro Data Structures, Discrete Math, Data Structures and Algorithms, Intro to Artificial Intelligence, Intro Data Science, Principle Econ I, Robotics Mechanisms, AI Based Market Disruption, Calculus I, Data-Oriented Programming, AI Ethics, Data Manipulation, Building Data Driven Apps, Entrepreneurial Marketing, Models of Social Information

Expected May 2026

Harvard Data Science Initiative

Certificate of Completion · Agentic AI: Contextualized and Applied

Expected March 2026

Experience

Data Science Engineer

LightBox

Spearheaded and deployed end-to-end NLP pipelines in Python to extract, structure, and validate commercial real estate documents, processing up to 100 files per day and eliminating 3–4 hours of manual review per document.
Designed and iterated on GPT-based grading and HUD-FMR benchmarking systems to standardize 500+ page documents into structured 50-point summaries, directly enabling faster and more informed business decisions.
Evaluated and improved model accuracy through iterative testing and reporting, achieving 90–95% accuracy on document extraction and summarization tasks.

New York, NY · May 2023 – Present

Data Science Analyst

Capital Z Partners

Supported the firm's founders on a $2B leveraged acquisition of an insurance services business, contributing to due diligence, capital raising, negotiations, and transaction documentation.
Worked with the placement agent to model and structure co-investment capital raises targeting institutional and high net worth investors.
Deal signed July 2022; closed January 2023; 70% recap dividend paid October 2024; currently valued at 5x cost.

New York, NY · May 2022 – Aug 2022

Student Founder

Code Central

Helped conceive the business and worked alongside the founders on business planning, fundraising strategy, and early hiring from the ground up.
Built and refreshed programming and AI curricula across Python, C++, JavaScript, and Robotics for students ages 8–17.
Helped tutor students and scale the business to 400+ members and ~$1M in annual revenue.

New York, NY / Las Vegas, NV · Jan 2017 – May 2022

Projects

TwentyTwenty — AI-Powered Stock & Crypto Research Platforms

FastAPI · PostgreSQL · Claude API · Streamlit · pandas · yfinance · CoinGecko

Built two full-stack quantitative research platforms computing RSI, MACD, Bollinger Bands, and composite signal scores across FastAPI / PostgreSQL backends, with backtesting engines supporting multiple strategies and full metrics (Sharpe, Sortino, max drawdown) and AI-generated analyst briefs.

ScoutingReport — AI-Powered MLB Scouting Platform

FastAPI · Supabase · Claude API · Streamlit · Plotly · MLB Stats API

Built a full-stack MLB analytics platform combining live player data with AI-generated scouting reports and game previews. Features a pitch location heatmap visualizer with handedness and pitch-type filters, a head-to-head matchup analyzer, career stat trends, and a today's games dashboard with probable pitcher breakdowns and AI game previews — all backed by a FastAPI + Supabase data layer.

Zookeeper — Graph Algorithms & Travelling Salesman

C++ · Graph Theory · MST · TSP · Branch & Bound

Implemented three modes of a graph optimization engine in C++. Built Prim's algorithm to compute a Minimum Spanning Tree across a constrained map with zone-based connectivity rules. Designed and implemented a fast TSP heuristic (nearest-neighbor insertion) to approximate optimal tours across tens of thousands of nodes in O(n²) time. Implemented an exact branch-and-bound TSP solver with MST-based lower bound pruning to guarantee optimal solutions within strict time limits.

Elevator Escape — C++ Elevator Simulation & AI

C++ · OOP · Greedy AI · Terminal Simulation

Built a 10-floor hotel elevator simulation in C++ with human and AI play modes. The AI uses an anger-weighted greedy dispatch algorithm — scoring every floor by cumulative passenger frustration, routing the closest idle elevator, and optimizing each pickup by loading whichever direction group (up vs. down) carries the higher combined anger to maximize satisfaction per trip. Engineered the full simulation engine: per-tick elevator movement, anger escalation, passenger explosions, a move-command parser, and save/load support.

LightBox Rent-Roll AI System

End-to-end ML pipeline · Python · GPT-based graders · HUD-FMR benchmarking

Built an end-to-end pipeline to extract, standardize, and validate rent-roll data using Python, GPT-based graders, and automated HUD-FMR benchmarking, reducing QA time by ~80% and powering interactive analytics dashboards.